initial
Browse files- README.md +53 -0
- config.json +28 -0
- merges.txt +0 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1816 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- ../../data/skquad
|
| 7 |
+
model-index:
|
| 8 |
+
- name: model
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# model
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [gerulata/slovakbert](https://huggingface.co/gerulata/slovakbert) on the ../../data/skquad dataset.
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 3e-05
|
| 37 |
+
- train_batch_size: 8
|
| 38 |
+
- eval_batch_size: 8
|
| 39 |
+
- seed: 42
|
| 40 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 41 |
+
- lr_scheduler_type: linear
|
| 42 |
+
- num_epochs: 3.0
|
| 43 |
+
|
| 44 |
+
### Training results
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
### Framework versions
|
| 49 |
+
|
| 50 |
+
- Transformers 4.16.0.dev0
|
| 51 |
+
- Pytorch 1.8.1
|
| 52 |
+
- Datasets 1.15.1
|
| 53 |
+
- Tokenizers 0.11.0
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "gerulata/slovakbert",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForQuestionAnswering"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"gradient_checkpointing": false,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_norm_eps": 1e-05,
|
| 17 |
+
"max_position_embeddings": 514,
|
| 18 |
+
"model_type": "roberta",
|
| 19 |
+
"num_attention_heads": 12,
|
| 20 |
+
"num_hidden_layers": 12,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"torch_dtype": "float32",
|
| 24 |
+
"transformers_version": "4.16.0.dev0",
|
| 25 |
+
"type_vocab_size": 1,
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 50264
|
| 28 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22d9944dd351c4896430ebb7a8a54dc60b04c77753cecc2a03f9b7bbd90a1791
|
| 3 |
+
size 992571373
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8cfb903c9714727b784b9bf38a1f8019e317ec6b3c39f8006d744e3e51735b5
|
| 3 |
+
size 496313015
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:413231525411e099a3027bf75976a8c9661c8320c27692b2f02c25827571afff
|
| 3 |
+
size 14657
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36903f490e9ca8bc06206f12b7dfb0b4d97c5a56f6e3b7e62df1c2d520afec9f
|
| 3 |
+
size 623
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true}}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "errors": "replace", "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "trim_offsets": true, "special_tokens_map_file": null, "model_max_length": 512, "name_or_path": "gerulata/slovakbert", "tokenizer_class": "RobertaTokenizer"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.8957528957528957,
|
| 5 |
+
"global_step": 30000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 2.9903474903474904e-05,
|
| 13 |
+
"loss": 4.7662,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 2.9806949806949808e-05,
|
| 19 |
+
"loss": 5.8747,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.03,
|
| 24 |
+
"learning_rate": 2.9710424710424712e-05,
|
| 25 |
+
"loss": 6.2395,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.04,
|
| 30 |
+
"learning_rate": 2.9613899613899616e-05,
|
| 31 |
+
"loss": 6.2412,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.05,
|
| 36 |
+
"learning_rate": 2.9517374517374516e-05,
|
| 37 |
+
"loss": 6.2405,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.06,
|
| 42 |
+
"learning_rate": 2.9420849420849423e-05,
|
| 43 |
+
"loss": 3.4061,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.07,
|
| 48 |
+
"learning_rate": 2.9324324324324327e-05,
|
| 49 |
+
"loss": 2.1295,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.08,
|
| 54 |
+
"learning_rate": 2.9227799227799227e-05,
|
| 55 |
+
"loss": 2.0931,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.09,
|
| 60 |
+
"learning_rate": 2.913127413127413e-05,
|
| 61 |
+
"loss": 1.9645,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.1,
|
| 66 |
+
"learning_rate": 2.9034749034749038e-05,
|
| 67 |
+
"loss": 1.8975,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.11,
|
| 72 |
+
"learning_rate": 2.8938223938223938e-05,
|
| 73 |
+
"loss": 1.8587,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.12,
|
| 78 |
+
"learning_rate": 2.8841698841698842e-05,
|
| 79 |
+
"loss": 1.7731,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.13,
|
| 84 |
+
"learning_rate": 2.874517374517375e-05,
|
| 85 |
+
"loss": 1.7233,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.14,
|
| 90 |
+
"learning_rate": 2.864864864864865e-05,
|
| 91 |
+
"loss": 1.6743,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.14,
|
| 96 |
+
"learning_rate": 2.8552123552123553e-05,
|
| 97 |
+
"loss": 1.6522,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.15,
|
| 102 |
+
"learning_rate": 2.8455598455598453e-05,
|
| 103 |
+
"loss": 1.579,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.16,
|
| 108 |
+
"learning_rate": 2.835907335907336e-05,
|
| 109 |
+
"loss": 1.6699,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.17,
|
| 114 |
+
"learning_rate": 2.8262548262548264e-05,
|
| 115 |
+
"loss": 2.5233,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.18,
|
| 120 |
+
"learning_rate": 2.8166023166023164e-05,
|
| 121 |
+
"loss": 1.8225,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.19,
|
| 126 |
+
"learning_rate": 2.806949806949807e-05,
|
| 127 |
+
"loss": 1.5727,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.2,
|
| 132 |
+
"learning_rate": 2.7972972972972975e-05,
|
| 133 |
+
"loss": 1.5446,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.21,
|
| 138 |
+
"learning_rate": 2.7876447876447876e-05,
|
| 139 |
+
"loss": 1.5248,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.22,
|
| 144 |
+
"learning_rate": 2.777992277992278e-05,
|
| 145 |
+
"loss": 1.5045,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.23,
|
| 150 |
+
"learning_rate": 2.7683397683397686e-05,
|
| 151 |
+
"loss": 1.4864,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.24,
|
| 156 |
+
"learning_rate": 2.7586872586872587e-05,
|
| 157 |
+
"loss": 1.4444,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.25,
|
| 162 |
+
"learning_rate": 2.749034749034749e-05,
|
| 163 |
+
"loss": 1.4808,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.26,
|
| 168 |
+
"learning_rate": 2.7393822393822394e-05,
|
| 169 |
+
"loss": 1.4867,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.27,
|
| 174 |
+
"learning_rate": 2.7297297297297298e-05,
|
| 175 |
+
"loss": 1.4764,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.28,
|
| 180 |
+
"learning_rate": 2.72007722007722e-05,
|
| 181 |
+
"loss": 1.4469,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.29,
|
| 186 |
+
"learning_rate": 2.7104247104247105e-05,
|
| 187 |
+
"loss": 1.4145,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.3,
|
| 192 |
+
"learning_rate": 2.700772200772201e-05,
|
| 193 |
+
"loss": 1.4232,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.31,
|
| 198 |
+
"learning_rate": 2.6911196911196913e-05,
|
| 199 |
+
"loss": 1.3993,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.32,
|
| 204 |
+
"learning_rate": 2.6814671814671816e-05,
|
| 205 |
+
"loss": 1.4332,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.33,
|
| 210 |
+
"learning_rate": 2.6718146718146717e-05,
|
| 211 |
+
"loss": 1.4679,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.34,
|
| 216 |
+
"learning_rate": 2.6621621621621624e-05,
|
| 217 |
+
"loss": 1.3816,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.35,
|
| 222 |
+
"learning_rate": 2.6525096525096527e-05,
|
| 223 |
+
"loss": 1.4694,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.36,
|
| 228 |
+
"learning_rate": 2.6428571428571428e-05,
|
| 229 |
+
"loss": 1.4303,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.37,
|
| 234 |
+
"learning_rate": 2.633204633204633e-05,
|
| 235 |
+
"loss": 1.3515,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.38,
|
| 240 |
+
"learning_rate": 2.623552123552124e-05,
|
| 241 |
+
"loss": 1.4127,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.39,
|
| 246 |
+
"learning_rate": 2.613899613899614e-05,
|
| 247 |
+
"loss": 1.3385,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.4,
|
| 252 |
+
"learning_rate": 2.6042471042471043e-05,
|
| 253 |
+
"loss": 1.2839,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.41,
|
| 258 |
+
"learning_rate": 2.594594594594595e-05,
|
| 259 |
+
"loss": 1.35,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.42,
|
| 264 |
+
"learning_rate": 2.584942084942085e-05,
|
| 265 |
+
"loss": 1.3536,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.42,
|
| 270 |
+
"learning_rate": 2.5752895752895754e-05,
|
| 271 |
+
"loss": 1.3542,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.43,
|
| 276 |
+
"learning_rate": 2.5656370656370657e-05,
|
| 277 |
+
"loss": 1.3096,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.44,
|
| 282 |
+
"learning_rate": 2.555984555984556e-05,
|
| 283 |
+
"loss": 1.4777,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.45,
|
| 288 |
+
"learning_rate": 2.5463320463320465e-05,
|
| 289 |
+
"loss": 1.2712,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.46,
|
| 294 |
+
"learning_rate": 2.5366795366795365e-05,
|
| 295 |
+
"loss": 1.3084,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.47,
|
| 300 |
+
"learning_rate": 2.527027027027027e-05,
|
| 301 |
+
"loss": 1.3296,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.48,
|
| 306 |
+
"learning_rate": 2.5173745173745176e-05,
|
| 307 |
+
"loss": 1.5364,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.49,
|
| 312 |
+
"learning_rate": 2.5077220077220076e-05,
|
| 313 |
+
"loss": 2.753,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.5,
|
| 318 |
+
"learning_rate": 2.498069498069498e-05,
|
| 319 |
+
"loss": 1.742,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.51,
|
| 324 |
+
"learning_rate": 2.4884169884169887e-05,
|
| 325 |
+
"loss": 1.6347,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.52,
|
| 330 |
+
"learning_rate": 2.4787644787644787e-05,
|
| 331 |
+
"loss": 1.6604,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.53,
|
| 336 |
+
"learning_rate": 2.469111969111969e-05,
|
| 337 |
+
"loss": 1.5615,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.54,
|
| 342 |
+
"learning_rate": 2.4594594594594595e-05,
|
| 343 |
+
"loss": 1.3292,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.55,
|
| 348 |
+
"learning_rate": 2.44980694980695e-05,
|
| 349 |
+
"loss": 1.2548,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.56,
|
| 354 |
+
"learning_rate": 2.4401544401544402e-05,
|
| 355 |
+
"loss": 1.2897,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.57,
|
| 360 |
+
"learning_rate": 2.4305019305019306e-05,
|
| 361 |
+
"loss": 1.2414,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.58,
|
| 366 |
+
"learning_rate": 2.420849420849421e-05,
|
| 367 |
+
"loss": 1.2883,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.59,
|
| 372 |
+
"learning_rate": 2.4111969111969113e-05,
|
| 373 |
+
"loss": 1.3017,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.6,
|
| 378 |
+
"learning_rate": 2.4015444015444017e-05,
|
| 379 |
+
"loss": 1.2273,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.61,
|
| 384 |
+
"learning_rate": 2.3918918918918917e-05,
|
| 385 |
+
"loss": 1.2292,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.62,
|
| 390 |
+
"learning_rate": 2.3822393822393824e-05,
|
| 391 |
+
"loss": 1.3544,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.63,
|
| 396 |
+
"learning_rate": 2.3725868725868728e-05,
|
| 397 |
+
"loss": 1.2666,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.64,
|
| 402 |
+
"learning_rate": 2.362934362934363e-05,
|
| 403 |
+
"loss": 1.2791,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.65,
|
| 408 |
+
"learning_rate": 2.3532818532818532e-05,
|
| 409 |
+
"loss": 1.3095,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.66,
|
| 414 |
+
"learning_rate": 2.343629343629344e-05,
|
| 415 |
+
"loss": 1.2471,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.67,
|
| 420 |
+
"learning_rate": 2.333976833976834e-05,
|
| 421 |
+
"loss": 1.3152,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.68,
|
| 426 |
+
"learning_rate": 2.3243243243243243e-05,
|
| 427 |
+
"loss": 1.1917,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.69,
|
| 432 |
+
"learning_rate": 2.314671814671815e-05,
|
| 433 |
+
"loss": 1.2383,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.69,
|
| 438 |
+
"learning_rate": 2.305019305019305e-05,
|
| 439 |
+
"loss": 1.2902,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.7,
|
| 444 |
+
"learning_rate": 2.2953667953667954e-05,
|
| 445 |
+
"loss": 1.2329,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.71,
|
| 450 |
+
"learning_rate": 2.2857142857142858e-05,
|
| 451 |
+
"loss": 1.1985,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.72,
|
| 456 |
+
"learning_rate": 2.2760617760617762e-05,
|
| 457 |
+
"loss": 1.1914,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.73,
|
| 462 |
+
"learning_rate": 2.2664092664092665e-05,
|
| 463 |
+
"loss": 1.2317,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.74,
|
| 468 |
+
"learning_rate": 2.2567567567567566e-05,
|
| 469 |
+
"loss": 1.2853,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.75,
|
| 474 |
+
"learning_rate": 2.247104247104247e-05,
|
| 475 |
+
"loss": 1.2564,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.76,
|
| 480 |
+
"learning_rate": 2.2374517374517377e-05,
|
| 481 |
+
"loss": 1.2579,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.77,
|
| 486 |
+
"learning_rate": 2.2277992277992277e-05,
|
| 487 |
+
"loss": 1.1445,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.78,
|
| 492 |
+
"learning_rate": 2.218146718146718e-05,
|
| 493 |
+
"loss": 1.1992,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.79,
|
| 498 |
+
"learning_rate": 2.2084942084942088e-05,
|
| 499 |
+
"loss": 1.1943,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.8,
|
| 504 |
+
"learning_rate": 2.1988416988416988e-05,
|
| 505 |
+
"loss": 1.1685,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.81,
|
| 510 |
+
"learning_rate": 2.1891891891891892e-05,
|
| 511 |
+
"loss": 1.1765,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.82,
|
| 516 |
+
"learning_rate": 2.1795366795366795e-05,
|
| 517 |
+
"loss": 1.168,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.83,
|
| 522 |
+
"learning_rate": 2.16988416988417e-05,
|
| 523 |
+
"loss": 1.2075,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.84,
|
| 528 |
+
"learning_rate": 2.1602316602316603e-05,
|
| 529 |
+
"loss": 1.1779,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.85,
|
| 534 |
+
"learning_rate": 2.1505791505791507e-05,
|
| 535 |
+
"loss": 1.1858,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.86,
|
| 540 |
+
"learning_rate": 2.140926640926641e-05,
|
| 541 |
+
"loss": 1.1828,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.87,
|
| 546 |
+
"learning_rate": 2.1312741312741314e-05,
|
| 547 |
+
"loss": 1.2031,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.88,
|
| 552 |
+
"learning_rate": 2.1216216216216218e-05,
|
| 553 |
+
"loss": 1.1677,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.89,
|
| 558 |
+
"learning_rate": 2.1119691119691118e-05,
|
| 559 |
+
"loss": 1.1525,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.9,
|
| 564 |
+
"learning_rate": 2.1023166023166025e-05,
|
| 565 |
+
"loss": 1.1876,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.91,
|
| 570 |
+
"learning_rate": 2.092664092664093e-05,
|
| 571 |
+
"loss": 1.1632,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.92,
|
| 576 |
+
"learning_rate": 2.083011583011583e-05,
|
| 577 |
+
"loss": 1.1599,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.93,
|
| 582 |
+
"learning_rate": 2.0733590733590733e-05,
|
| 583 |
+
"loss": 1.1763,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.94,
|
| 588 |
+
"learning_rate": 2.063706563706564e-05,
|
| 589 |
+
"loss": 1.1605,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.95,
|
| 594 |
+
"learning_rate": 2.054054054054054e-05,
|
| 595 |
+
"loss": 1.3013,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.96,
|
| 600 |
+
"learning_rate": 2.0444015444015444e-05,
|
| 601 |
+
"loss": 1.1579,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.97,
|
| 606 |
+
"learning_rate": 2.034749034749035e-05,
|
| 607 |
+
"loss": 1.2543,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.97,
|
| 612 |
+
"learning_rate": 2.025096525096525e-05,
|
| 613 |
+
"loss": 1.1719,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.98,
|
| 618 |
+
"learning_rate": 2.0154440154440155e-05,
|
| 619 |
+
"loss": 1.1845,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.99,
|
| 624 |
+
"learning_rate": 2.005791505791506e-05,
|
| 625 |
+
"loss": 1.2079,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.0,
|
| 630 |
+
"learning_rate": 1.9961389961389962e-05,
|
| 631 |
+
"loss": 1.0919,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.01,
|
| 636 |
+
"learning_rate": 1.9864864864864866e-05,
|
| 637 |
+
"loss": 1.0697,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.02,
|
| 642 |
+
"learning_rate": 1.976833976833977e-05,
|
| 643 |
+
"loss": 1.009,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.03,
|
| 648 |
+
"learning_rate": 1.967181467181467e-05,
|
| 649 |
+
"loss": 1.0711,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.04,
|
| 654 |
+
"learning_rate": 1.9575289575289577e-05,
|
| 655 |
+
"loss": 0.9108,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.05,
|
| 660 |
+
"learning_rate": 1.9478764478764478e-05,
|
| 661 |
+
"loss": 0.9708,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.06,
|
| 666 |
+
"learning_rate": 1.938223938223938e-05,
|
| 667 |
+
"loss": 0.9146,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.07,
|
| 672 |
+
"learning_rate": 1.928571428571429e-05,
|
| 673 |
+
"loss": 0.9609,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.08,
|
| 678 |
+
"learning_rate": 1.918918918918919e-05,
|
| 679 |
+
"loss": 1.0282,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.09,
|
| 684 |
+
"learning_rate": 1.9092664092664092e-05,
|
| 685 |
+
"loss": 0.9474,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.1,
|
| 690 |
+
"learning_rate": 1.8996138996138996e-05,
|
| 691 |
+
"loss": 0.9342,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.11,
|
| 696 |
+
"learning_rate": 1.88996138996139e-05,
|
| 697 |
+
"loss": 0.9938,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.12,
|
| 702 |
+
"learning_rate": 1.8803088803088804e-05,
|
| 703 |
+
"loss": 0.9652,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.13,
|
| 708 |
+
"learning_rate": 1.8706563706563707e-05,
|
| 709 |
+
"loss": 0.9364,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.14,
|
| 714 |
+
"learning_rate": 1.8610038610038608e-05,
|
| 715 |
+
"loss": 0.9736,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.15,
|
| 720 |
+
"learning_rate": 1.8513513513513515e-05,
|
| 721 |
+
"loss": 1.0112,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.16,
|
| 726 |
+
"learning_rate": 1.841698841698842e-05,
|
| 727 |
+
"loss": 1.0339,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.17,
|
| 732 |
+
"learning_rate": 1.832046332046332e-05,
|
| 733 |
+
"loss": 0.9825,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.18,
|
| 738 |
+
"learning_rate": 1.8223938223938226e-05,
|
| 739 |
+
"loss": 0.9595,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.19,
|
| 744 |
+
"learning_rate": 1.812741312741313e-05,
|
| 745 |
+
"loss": 0.9858,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.2,
|
| 750 |
+
"learning_rate": 1.803088803088803e-05,
|
| 751 |
+
"loss": 0.9754,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.21,
|
| 756 |
+
"learning_rate": 1.7934362934362933e-05,
|
| 757 |
+
"loss": 1.0461,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.22,
|
| 762 |
+
"learning_rate": 1.783783783783784e-05,
|
| 763 |
+
"loss": 0.9761,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.23,
|
| 768 |
+
"learning_rate": 1.774131274131274e-05,
|
| 769 |
+
"loss": 0.999,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.24,
|
| 774 |
+
"learning_rate": 1.7644787644787645e-05,
|
| 775 |
+
"loss": 0.9608,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.25,
|
| 780 |
+
"learning_rate": 1.7548262548262552e-05,
|
| 781 |
+
"loss": 0.9852,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.25,
|
| 786 |
+
"learning_rate": 1.7451737451737452e-05,
|
| 787 |
+
"loss": 0.9841,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.26,
|
| 792 |
+
"learning_rate": 1.7355212355212356e-05,
|
| 793 |
+
"loss": 0.9992,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.27,
|
| 798 |
+
"learning_rate": 1.725868725868726e-05,
|
| 799 |
+
"loss": 0.9912,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.28,
|
| 804 |
+
"learning_rate": 1.7162162162162163e-05,
|
| 805 |
+
"loss": 0.998,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.29,
|
| 810 |
+
"learning_rate": 1.7065637065637067e-05,
|
| 811 |
+
"loss": 0.9671,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.3,
|
| 816 |
+
"learning_rate": 1.696911196911197e-05,
|
| 817 |
+
"loss": 0.9372,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.31,
|
| 822 |
+
"learning_rate": 1.687258687258687e-05,
|
| 823 |
+
"loss": 1.0567,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.32,
|
| 828 |
+
"learning_rate": 1.6776061776061778e-05,
|
| 829 |
+
"loss": 0.9795,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.33,
|
| 834 |
+
"learning_rate": 1.6679536679536678e-05,
|
| 835 |
+
"loss": 0.955,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.34,
|
| 840 |
+
"learning_rate": 1.6583011583011582e-05,
|
| 841 |
+
"loss": 1.0127,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.35,
|
| 846 |
+
"learning_rate": 1.648648648648649e-05,
|
| 847 |
+
"loss": 0.9543,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.36,
|
| 852 |
+
"learning_rate": 1.638996138996139e-05,
|
| 853 |
+
"loss": 0.9838,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.37,
|
| 858 |
+
"learning_rate": 1.6293436293436293e-05,
|
| 859 |
+
"loss": 0.9507,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.38,
|
| 864 |
+
"learning_rate": 1.6196911196911197e-05,
|
| 865 |
+
"loss": 0.9387,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.39,
|
| 870 |
+
"learning_rate": 1.61003861003861e-05,
|
| 871 |
+
"loss": 0.9229,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.4,
|
| 876 |
+
"learning_rate": 1.6003861003861004e-05,
|
| 877 |
+
"loss": 0.99,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.41,
|
| 882 |
+
"learning_rate": 1.5907335907335908e-05,
|
| 883 |
+
"loss": 0.9069,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.42,
|
| 888 |
+
"learning_rate": 1.5810810810810808e-05,
|
| 889 |
+
"loss": 0.9713,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.43,
|
| 894 |
+
"learning_rate": 1.5714285714285715e-05,
|
| 895 |
+
"loss": 0.9303,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.44,
|
| 900 |
+
"learning_rate": 1.561776061776062e-05,
|
| 901 |
+
"loss": 0.9915,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.45,
|
| 906 |
+
"learning_rate": 1.552123552123552e-05,
|
| 907 |
+
"loss": 0.9553,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 1.46,
|
| 912 |
+
"learning_rate": 1.5424710424710426e-05,
|
| 913 |
+
"loss": 0.9928,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.47,
|
| 918 |
+
"learning_rate": 1.532818532818533e-05,
|
| 919 |
+
"loss": 0.9777,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.48,
|
| 924 |
+
"learning_rate": 1.5231660231660232e-05,
|
| 925 |
+
"loss": 0.9512,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.49,
|
| 930 |
+
"learning_rate": 1.5135135135135134e-05,
|
| 931 |
+
"loss": 0.9259,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.5,
|
| 936 |
+
"learning_rate": 1.503861003861004e-05,
|
| 937 |
+
"loss": 0.8975,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.51,
|
| 942 |
+
"learning_rate": 1.4942084942084943e-05,
|
| 943 |
+
"loss": 1.0091,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.52,
|
| 948 |
+
"learning_rate": 1.4845559845559845e-05,
|
| 949 |
+
"loss": 1.0214,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 1.53,
|
| 954 |
+
"learning_rate": 1.4749034749034749e-05,
|
| 955 |
+
"loss": 0.9119,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.53,
|
| 960 |
+
"learning_rate": 1.4652509652509654e-05,
|
| 961 |
+
"loss": 0.9976,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 1.54,
|
| 966 |
+
"learning_rate": 1.4555984555984556e-05,
|
| 967 |
+
"loss": 0.9755,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 1.55,
|
| 972 |
+
"learning_rate": 1.445945945945946e-05,
|
| 973 |
+
"loss": 0.9176,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.56,
|
| 978 |
+
"learning_rate": 1.4362934362934362e-05,
|
| 979 |
+
"loss": 0.9487,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 1.57,
|
| 984 |
+
"learning_rate": 1.4266409266409268e-05,
|
| 985 |
+
"loss": 0.947,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 1.58,
|
| 990 |
+
"learning_rate": 1.416988416988417e-05,
|
| 991 |
+
"loss": 1.0008,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 1.59,
|
| 996 |
+
"learning_rate": 1.4073359073359073e-05,
|
| 997 |
+
"loss": 0.9436,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.6,
|
| 1002 |
+
"learning_rate": 1.3976833976833977e-05,
|
| 1003 |
+
"loss": 0.9633,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.61,
|
| 1008 |
+
"learning_rate": 1.388030888030888e-05,
|
| 1009 |
+
"loss": 0.9685,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 1.62,
|
| 1014 |
+
"learning_rate": 1.3783783783783784e-05,
|
| 1015 |
+
"loss": 0.9424,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.63,
|
| 1020 |
+
"learning_rate": 1.3687258687258688e-05,
|
| 1021 |
+
"loss": 0.8993,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 1.64,
|
| 1026 |
+
"learning_rate": 1.3590733590733592e-05,
|
| 1027 |
+
"loss": 0.9622,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 1.65,
|
| 1032 |
+
"learning_rate": 1.3494208494208494e-05,
|
| 1033 |
+
"loss": 0.9735,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 1.66,
|
| 1038 |
+
"learning_rate": 1.33976833976834e-05,
|
| 1039 |
+
"loss": 0.9564,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 1.67,
|
| 1044 |
+
"learning_rate": 1.3301158301158301e-05,
|
| 1045 |
+
"loss": 0.9575,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 1.68,
|
| 1050 |
+
"learning_rate": 1.3204633204633205e-05,
|
| 1051 |
+
"loss": 0.8914,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 1.69,
|
| 1056 |
+
"learning_rate": 1.3108108108108109e-05,
|
| 1057 |
+
"loss": 0.9441,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.7,
|
| 1062 |
+
"learning_rate": 1.3011583011583012e-05,
|
| 1063 |
+
"loss": 1.0149,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 1.71,
|
| 1068 |
+
"learning_rate": 1.2915057915057914e-05,
|
| 1069 |
+
"loss": 0.9439,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 1.72,
|
| 1074 |
+
"learning_rate": 1.2818532818532818e-05,
|
| 1075 |
+
"loss": 0.9569,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 1.73,
|
| 1080 |
+
"learning_rate": 1.2722007722007723e-05,
|
| 1081 |
+
"loss": 0.8933,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 1.74,
|
| 1086 |
+
"learning_rate": 1.2625482625482625e-05,
|
| 1087 |
+
"loss": 0.9097,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.75,
|
| 1092 |
+
"learning_rate": 1.2528957528957529e-05,
|
| 1093 |
+
"loss": 0.8836,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.76,
|
| 1098 |
+
"learning_rate": 1.2432432432432433e-05,
|
| 1099 |
+
"loss": 0.9698,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.77,
|
| 1104 |
+
"learning_rate": 1.2335907335907337e-05,
|
| 1105 |
+
"loss": 0.9441,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 1.78,
|
| 1110 |
+
"learning_rate": 1.2239382239382239e-05,
|
| 1111 |
+
"loss": 0.9384,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 1.79,
|
| 1116 |
+
"learning_rate": 1.2142857142857144e-05,
|
| 1117 |
+
"loss": 0.9416,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 1.8,
|
| 1122 |
+
"learning_rate": 1.2046332046332046e-05,
|
| 1123 |
+
"loss": 0.9299,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.81,
|
| 1128 |
+
"learning_rate": 1.194980694980695e-05,
|
| 1129 |
+
"loss": 0.967,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.81,
|
| 1134 |
+
"learning_rate": 1.1853281853281855e-05,
|
| 1135 |
+
"loss": 0.8969,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.82,
|
| 1140 |
+
"learning_rate": 1.1756756756756757e-05,
|
| 1141 |
+
"loss": 0.925,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.83,
|
| 1146 |
+
"learning_rate": 1.166023166023166e-05,
|
| 1147 |
+
"loss": 0.942,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 1.84,
|
| 1152 |
+
"learning_rate": 1.1563706563706563e-05,
|
| 1153 |
+
"loss": 0.982,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 1.85,
|
| 1158 |
+
"learning_rate": 1.1467181467181468e-05,
|
| 1159 |
+
"loss": 0.9224,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 1.86,
|
| 1164 |
+
"learning_rate": 1.137065637065637e-05,
|
| 1165 |
+
"loss": 0.9527,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 1.87,
|
| 1170 |
+
"learning_rate": 1.1274131274131274e-05,
|
| 1171 |
+
"loss": 0.9188,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.88,
|
| 1176 |
+
"learning_rate": 1.1177606177606178e-05,
|
| 1177 |
+
"loss": 0.9603,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.89,
|
| 1182 |
+
"learning_rate": 1.1081081081081081e-05,
|
| 1183 |
+
"loss": 0.9158,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 1.9,
|
| 1188 |
+
"learning_rate": 1.0984555984555985e-05,
|
| 1189 |
+
"loss": 0.9615,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 1.91,
|
| 1194 |
+
"learning_rate": 1.0888030888030889e-05,
|
| 1195 |
+
"loss": 0.9307,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 1.92,
|
| 1200 |
+
"learning_rate": 1.0791505791505792e-05,
|
| 1201 |
+
"loss": 0.8881,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 1.93,
|
| 1206 |
+
"learning_rate": 1.0694980694980694e-05,
|
| 1207 |
+
"loss": 0.8812,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 1.94,
|
| 1212 |
+
"learning_rate": 1.05984555984556e-05,
|
| 1213 |
+
"loss": 0.9301,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.95,
|
| 1218 |
+
"learning_rate": 1.0501930501930502e-05,
|
| 1219 |
+
"loss": 0.9501,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.96,
|
| 1224 |
+
"learning_rate": 1.0405405405405406e-05,
|
| 1225 |
+
"loss": 0.9434,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.97,
|
| 1230 |
+
"learning_rate": 1.030888030888031e-05,
|
| 1231 |
+
"loss": 0.9466,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 1.98,
|
| 1236 |
+
"learning_rate": 1.0212355212355213e-05,
|
| 1237 |
+
"loss": 0.8856,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 1.99,
|
| 1242 |
+
"learning_rate": 1.0115830115830115e-05,
|
| 1243 |
+
"loss": 0.9773,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.0,
|
| 1248 |
+
"learning_rate": 1.0019305019305019e-05,
|
| 1249 |
+
"loss": 0.8732,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 2.01,
|
| 1254 |
+
"learning_rate": 9.922779922779924e-06,
|
| 1255 |
+
"loss": 0.7991,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.02,
|
| 1260 |
+
"learning_rate": 9.826254826254826e-06,
|
| 1261 |
+
"loss": 0.7911,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.03,
|
| 1266 |
+
"learning_rate": 9.72972972972973e-06,
|
| 1267 |
+
"loss": 0.7547,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 2.04,
|
| 1272 |
+
"learning_rate": 9.633204633204634e-06,
|
| 1273 |
+
"loss": 0.7243,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.05,
|
| 1278 |
+
"learning_rate": 9.536679536679537e-06,
|
| 1279 |
+
"loss": 0.7338,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.06,
|
| 1284 |
+
"learning_rate": 9.44015444015444e-06,
|
| 1285 |
+
"loss": 0.7552,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.07,
|
| 1290 |
+
"learning_rate": 9.343629343629345e-06,
|
| 1291 |
+
"loss": 0.7639,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 2.08,
|
| 1296 |
+
"learning_rate": 9.247104247104247e-06,
|
| 1297 |
+
"loss": 0.7705,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 2.08,
|
| 1302 |
+
"learning_rate": 9.15057915057915e-06,
|
| 1303 |
+
"loss": 0.7605,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.09,
|
| 1308 |
+
"learning_rate": 9.054054054054056e-06,
|
| 1309 |
+
"loss": 0.7734,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 2.1,
|
| 1314 |
+
"learning_rate": 8.957528957528958e-06,
|
| 1315 |
+
"loss": 0.747,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 2.11,
|
| 1320 |
+
"learning_rate": 8.861003861003861e-06,
|
| 1321 |
+
"loss": 0.7564,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.12,
|
| 1326 |
+
"learning_rate": 8.764478764478765e-06,
|
| 1327 |
+
"loss": 0.7362,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 2.13,
|
| 1332 |
+
"learning_rate": 8.667953667953669e-06,
|
| 1333 |
+
"loss": 0.7052,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 2.14,
|
| 1338 |
+
"learning_rate": 8.571428571428571e-06,
|
| 1339 |
+
"loss": 0.7525,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 2.15,
|
| 1344 |
+
"learning_rate": 8.474903474903475e-06,
|
| 1345 |
+
"loss": 0.754,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 2.16,
|
| 1350 |
+
"learning_rate": 8.378378378378378e-06,
|
| 1351 |
+
"loss": 0.7815,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 2.17,
|
| 1356 |
+
"learning_rate": 8.281853281853282e-06,
|
| 1357 |
+
"loss": 0.7356,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 2.18,
|
| 1362 |
+
"learning_rate": 8.185328185328184e-06,
|
| 1363 |
+
"loss": 0.7302,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.19,
|
| 1368 |
+
"learning_rate": 8.08880308880309e-06,
|
| 1369 |
+
"loss": 0.7673,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 2.2,
|
| 1374 |
+
"learning_rate": 7.992277992277993e-06,
|
| 1375 |
+
"loss": 0.7557,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 2.21,
|
| 1380 |
+
"learning_rate": 7.895752895752895e-06,
|
| 1381 |
+
"loss": 0.7523,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.22,
|
| 1386 |
+
"learning_rate": 7.7992277992278e-06,
|
| 1387 |
+
"loss": 0.7761,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 2.23,
|
| 1392 |
+
"learning_rate": 7.702702702702703e-06,
|
| 1393 |
+
"loss": 0.7202,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 2.24,
|
| 1398 |
+
"learning_rate": 7.606177606177607e-06,
|
| 1399 |
+
"loss": 0.7508,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 2.25,
|
| 1404 |
+
"learning_rate": 7.509652509652509e-06,
|
| 1405 |
+
"loss": 0.7161,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.26,
|
| 1410 |
+
"learning_rate": 7.413127413127414e-06,
|
| 1411 |
+
"loss": 0.8187,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 2.27,
|
| 1416 |
+
"learning_rate": 7.3166023166023166e-06,
|
| 1417 |
+
"loss": 0.7501,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 2.28,
|
| 1422 |
+
"learning_rate": 7.22007722007722e-06,
|
| 1423 |
+
"loss": 0.7494,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 2.29,
|
| 1428 |
+
"learning_rate": 7.123552123552124e-06,
|
| 1429 |
+
"loss": 0.7742,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 2.3,
|
| 1434 |
+
"learning_rate": 7.027027027027027e-06,
|
| 1435 |
+
"loss": 0.7279,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 2.31,
|
| 1440 |
+
"learning_rate": 6.9305019305019305e-06,
|
| 1441 |
+
"loss": 0.7401,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.32,
|
| 1446 |
+
"learning_rate": 6.833976833976834e-06,
|
| 1447 |
+
"loss": 0.7151,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.33,
|
| 1452 |
+
"learning_rate": 6.737451737451738e-06,
|
| 1453 |
+
"loss": 0.7656,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 2.34,
|
| 1458 |
+
"learning_rate": 6.640926640926642e-06,
|
| 1459 |
+
"loss": 0.7402,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 2.35,
|
| 1464 |
+
"learning_rate": 6.5444015444015445e-06,
|
| 1465 |
+
"loss": 0.7264,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 2.36,
|
| 1470 |
+
"learning_rate": 6.447876447876448e-06,
|
| 1471 |
+
"loss": 0.7397,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 2.36,
|
| 1476 |
+
"learning_rate": 6.351351351351352e-06,
|
| 1477 |
+
"loss": 0.7495,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 2.37,
|
| 1482 |
+
"learning_rate": 6.254826254826255e-06,
|
| 1483 |
+
"loss": 0.7814,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 2.38,
|
| 1488 |
+
"learning_rate": 6.1583011583011585e-06,
|
| 1489 |
+
"loss": 0.7627,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.39,
|
| 1494 |
+
"learning_rate": 6.061776061776061e-06,
|
| 1495 |
+
"loss": 0.6907,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 2.4,
|
| 1500 |
+
"learning_rate": 5.965250965250965e-06,
|
| 1501 |
+
"loss": 0.7289,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 2.41,
|
| 1506 |
+
"learning_rate": 5.868725868725869e-06,
|
| 1507 |
+
"loss": 0.7333,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 2.42,
|
| 1512 |
+
"learning_rate": 5.7722007722007725e-06,
|
| 1513 |
+
"loss": 0.7544,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 2.43,
|
| 1518 |
+
"learning_rate": 5.675675675675676e-06,
|
| 1519 |
+
"loss": 0.753,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 2.44,
|
| 1524 |
+
"learning_rate": 5.57915057915058e-06,
|
| 1525 |
+
"loss": 0.7658,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 2.45,
|
| 1530 |
+
"learning_rate": 5.482625482625483e-06,
|
| 1531 |
+
"loss": 0.7222,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.46,
|
| 1536 |
+
"learning_rate": 5.3861003861003864e-06,
|
| 1537 |
+
"loss": 0.726,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 2.47,
|
| 1542 |
+
"learning_rate": 5.289575289575289e-06,
|
| 1543 |
+
"loss": 0.7128,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 2.48,
|
| 1548 |
+
"learning_rate": 5.193050193050193e-06,
|
| 1549 |
+
"loss": 0.7381,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 2.49,
|
| 1554 |
+
"learning_rate": 5.096525096525097e-06,
|
| 1555 |
+
"loss": 0.7035,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 2.5,
|
| 1560 |
+
"learning_rate": 4.9999999999999996e-06,
|
| 1561 |
+
"loss": 0.681,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 2.51,
|
| 1566 |
+
"learning_rate": 4.903474903474903e-06,
|
| 1567 |
+
"loss": 0.7638,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 2.52,
|
| 1572 |
+
"learning_rate": 4.806949806949808e-06,
|
| 1573 |
+
"loss": 0.7013,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 2.53,
|
| 1578 |
+
"learning_rate": 4.710424710424711e-06,
|
| 1579 |
+
"loss": 0.6841,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 2.54,
|
| 1584 |
+
"learning_rate": 4.613899613899614e-06,
|
| 1585 |
+
"loss": 0.753,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 2.55,
|
| 1590 |
+
"learning_rate": 4.517374517374517e-06,
|
| 1591 |
+
"loss": 0.7082,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 2.56,
|
| 1596 |
+
"learning_rate": 4.420849420849421e-06,
|
| 1597 |
+
"loss": 0.7213,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 2.57,
|
| 1602 |
+
"learning_rate": 4.324324324324325e-06,
|
| 1603 |
+
"loss": 0.7049,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 2.58,
|
| 1608 |
+
"learning_rate": 4.2277992277992275e-06,
|
| 1609 |
+
"loss": 0.6806,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 2.59,
|
| 1614 |
+
"learning_rate": 4.131274131274131e-06,
|
| 1615 |
+
"loss": 0.716,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 2.6,
|
| 1620 |
+
"learning_rate": 4.034749034749035e-06,
|
| 1621 |
+
"loss": 0.7481,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 2.61,
|
| 1626 |
+
"learning_rate": 3.938223938223938e-06,
|
| 1627 |
+
"loss": 0.7112,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 2.62,
|
| 1632 |
+
"learning_rate": 3.841698841698842e-06,
|
| 1633 |
+
"loss": 0.7514,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 2.63,
|
| 1638 |
+
"learning_rate": 3.745173745173745e-06,
|
| 1639 |
+
"loss": 0.6975,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 2.64,
|
| 1644 |
+
"learning_rate": 3.648648648648649e-06,
|
| 1645 |
+
"loss": 0.7166,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 2.64,
|
| 1650 |
+
"learning_rate": 3.552123552123552e-06,
|
| 1651 |
+
"loss": 0.7702,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 2.65,
|
| 1656 |
+
"learning_rate": 3.455598455598456e-06,
|
| 1657 |
+
"loss": 0.713,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 2.66,
|
| 1662 |
+
"learning_rate": 3.359073359073359e-06,
|
| 1663 |
+
"loss": 0.7266,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 2.67,
|
| 1668 |
+
"learning_rate": 3.2625482625482625e-06,
|
| 1669 |
+
"loss": 0.7089,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 2.68,
|
| 1674 |
+
"learning_rate": 3.166023166023166e-06,
|
| 1675 |
+
"loss": 0.6851,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 2.69,
|
| 1680 |
+
"learning_rate": 3.06949806949807e-06,
|
| 1681 |
+
"loss": 0.7234,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 2.7,
|
| 1686 |
+
"learning_rate": 2.972972972972973e-06,
|
| 1687 |
+
"loss": 0.6683,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 2.71,
|
| 1692 |
+
"learning_rate": 2.8764478764478764e-06,
|
| 1693 |
+
"loss": 0.7244,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 2.72,
|
| 1698 |
+
"learning_rate": 2.7799227799227797e-06,
|
| 1699 |
+
"loss": 0.7578,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 2.73,
|
| 1704 |
+
"learning_rate": 2.683397683397684e-06,
|
| 1705 |
+
"loss": 0.7604,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 2.74,
|
| 1710 |
+
"learning_rate": 2.586872586872587e-06,
|
| 1711 |
+
"loss": 0.7131,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 2.75,
|
| 1716 |
+
"learning_rate": 2.4903474903474904e-06,
|
| 1717 |
+
"loss": 0.7049,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.76,
|
| 1722 |
+
"learning_rate": 2.3938223938223937e-06,
|
| 1723 |
+
"loss": 0.7552,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 2.77,
|
| 1728 |
+
"learning_rate": 2.2972972972972974e-06,
|
| 1729 |
+
"loss": 0.6907,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 2.78,
|
| 1734 |
+
"learning_rate": 2.200772200772201e-06,
|
| 1735 |
+
"loss": 0.7238,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 2.79,
|
| 1740 |
+
"learning_rate": 2.1042471042471044e-06,
|
| 1741 |
+
"loss": 0.7431,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 2.8,
|
| 1746 |
+
"learning_rate": 2.0077220077220077e-06,
|
| 1747 |
+
"loss": 0.7096,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 2.81,
|
| 1752 |
+
"learning_rate": 1.9111969111969114e-06,
|
| 1753 |
+
"loss": 0.6985,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 2.82,
|
| 1758 |
+
"learning_rate": 1.8146718146718149e-06,
|
| 1759 |
+
"loss": 0.7541,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.83,
|
| 1764 |
+
"learning_rate": 1.7181467181467181e-06,
|
| 1765 |
+
"loss": 0.762,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 2.84,
|
| 1770 |
+
"learning_rate": 1.6216216216216219e-06,
|
| 1771 |
+
"loss": 0.6741,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 2.85,
|
| 1776 |
+
"learning_rate": 1.5250965250965251e-06,
|
| 1777 |
+
"loss": 0.6974,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 2.86,
|
| 1782 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 1783 |
+
"loss": 0.7249,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 2.87,
|
| 1788 |
+
"learning_rate": 1.3320463320463321e-06,
|
| 1789 |
+
"loss": 0.6886,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 2.88,
|
| 1794 |
+
"learning_rate": 1.2355212355212356e-06,
|
| 1795 |
+
"loss": 0.6499,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 2.89,
|
| 1800 |
+
"learning_rate": 1.1389961389961391e-06,
|
| 1801 |
+
"loss": 0.7713,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.9,
|
| 1806 |
+
"learning_rate": 1.0424710424710424e-06,
|
| 1807 |
+
"loss": 0.711,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
}
|
| 1810 |
+
],
|
| 1811 |
+
"max_steps": 31080,
|
| 1812 |
+
"num_train_epochs": 3,
|
| 1813 |
+
"total_flos": 6.270808605555917e+16,
|
| 1814 |
+
"trial_name": null,
|
| 1815 |
+
"trial_params": null
|
| 1816 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5addf778fc7aa6c3a7e4b1707beb32a97b03c7545b6358370e12bf52417bbf50
|
| 3 |
+
size 2927
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|